Assembly Language for Beginners

(Jeff_L) #1

1.15. SWITCH()/CASE/DEFAULT


But if the value ofais less or equals to 4, then it gets multiplied by 4 and added with the$LN11@ftable
address. That is how an address inside the table is constructed, pointing exactly to the element we need.
For example, let’s sayais equal to 2. 2 ∗4 = 8(all table elements are addresses in a 32-bit process and
that is why all elements are 4 bytes wide). The address of the$LN11@ftable + 8 is the table element
where the$LN4@flabel is stored.JMPfetches the$LN4@faddress from the table and jumps to it.


This table is sometimes calledjumptableorbranch table^94.


Then the correspondingprintf()is called with argument'two'.
Literally, thejmp DWORD PTR $LN11@f[ecx4]instruction impliesjump to the DWORD that is stored at
address$LN11@f + ecx
4.


npad(.1.7 on page 1038) is an assembly language macro that align the next label so that it will be stored
at an address aligned on a 4 bytes (or 16 bytes) boundary. This is very suitable for the processor since
it is able to fetch 32-bit values from memory through the memory bus, cache memory, etc., in a more
effective way if it is aligned.


(^94) The whole method was once calledcomputed GOTOin early versions of Fortran:wikipedia. Not quite relevant these days, but
what a term!

Free download pdf